Yulhwa Kim

Assistant Professor, Sungkyunkwan University

1 paper at NeurIPS 2025

Homepage· OpenReview· Semantic Scholar· Google Scholar

Poster Session 1

1 paper

Wednesday, December 3, 2025 · 11:00 AM → 2:00 PM

Exhibit Hall C,D,E

Reasoning Path Compression: Compressing Generation Trajectories for Efficient LLM Reasoning

#912 · Jiwon Song, Dongwon Jo, Yulhwa Kim, Jae-Joon Kim

Reasoning Path Compression accelerates inference of reasoning LLMs by periodically compressing KV cache of generated tokens exploiting their semantic sparsity.